# A fun game (for problem composers, mostly...)

A game leading to a given position is called a proof game. Of course any proof game can be prepended with 1.Nf3 Nf6 2.Ng1 Ng8, so we defined a shortest proof game as, duh, the shortest possible proof game. Especially interesting are unique shortest proof games. For example, 1.e4 e5 2.Nf3 is not unique - White could swap moves. In contrast, 1.e4 e5 2.Bc4 is unique.

Rules of the fun game: First one creating a position without a unique shortest proof game loses. And that's it.

Do you have ideas for strategies or even a forced win for one side? Checking might be a good idea to force the hand of the opponent; 1.e4 d5?! 2.Bb5+ and now Nc6 or c6 lose, but anything to d7 is OK. (EDIT: And don't let your opponent capture, since one-side noncommuting moves run out quite fast.)

(For testers: Proof game checkers exist.)

• Can you also lose by being checkmated? Oct 23, 2022 at 1:03
• @NoamD.Elkies: The problem forum where I found this game said nothing about that. I think any move ending the game (stalemate, 50 move rule, etc.) should be treated as an attempt to bail out :-) and thus lose. Oct 23, 2022 at 8:15
• @MiguelAmbrona Why not? Prepending 1 Nf3 Nh6 2 Ng1 Ng8 doesn't lose any castling rights. Oct 23, 2022 at 21:45
• @supercat: You can do a PDB search by going on pdb.dieschwalbe.de/index.jsp and enter K='unique proof game' AND SOL='#'. This is no surefire answer, as '#' might be omitted and not all UPGs are in the database. A fleeting search gave the P0002187 with whopping 38.5 moves! (418 games are shown, so even that might not be the record...) Oct 25, 2022 at 7:45
• @NoamD.Elkies (and HaukeReddmann): If the rules are that checkmate has the same result as in ordinary chess, then one interesting result is that 1. f3 is a losing move for white. Black has a couple losing moves, but the best response of e5 forces Kf2 and black can continue so that the whole game is forced as follows: 1. f3? e5 2. Kf2 Qh4+ 3. Ke3 Qd4#! Oct 27, 2022 at 1:16